Hallowed be the vegan! But, do vegans have a higher risk of stroke?

It's on the news everywhere right now that vegetarians apparently have a lower risk of heart attack - which is nothing new - but, according to a new study, vegetarians still have a lower risk of heart attack, BUT at the same time a higher risk of stroke.

Spoiler: Take-home message: Vitamin B12 deficiency can likely increase the risk of stroke. Make sure you get enough vitamin B12. More info: Nutrient recommendations for vegans

Now, this isn't just any study, but this is the "EPIC-Oxford study" - one of the two big studies that contain a large number of vegans (the other one being the Adventist Health Study II = AHS2).

The main article that news media are reporting about contains a comparison of three different dietary groups:
- somewhat health-conscious non-vegetarians ("meat eaters")
- pescatarians (no meat, other than fish, possibly also other aquatic animals)
- vegetarians (including a smaller percentage, about 11%, of vegans)

The participants were recruited from 1993 to 2001 from all over the United Kingdom and categorized by diet. So the vegans in this study were already vegan then. The participants were again asked about their diet again in 2010 (up to 2013). So the vegans in this study were also vegan in ~2010 - but this is not only clear for about 57% of the vegans because only 43% of the vegans replied the second time they were asked (2010 to 2013; "28 364 participants returned a follow-up questionnaire in 2010 to 2013, including: [...] 792 vegans" (Tong et al. 2019 supplementary PDF)).
Can you believe it? Vegans who don't want to TELL you that they are vegan!
Disease events, including strokes and heart attacks in any of the participants, were recorded until 2016 (Tong et al. 2019).

Now, there are some people who like to say that whenever vegans get ill it's because they are "JUNK FOOD VEGANS!" - this is not true.
In this study, "vegetarians had higher intakes of fruit and vegetables, legumes and soya foods, nuts, and dietary fibre, and had a lower intake of saturated fat (10% v 12% energy) and sodium than meat eaters." (Tong et al. 2019)
What really interests us (me) is vegans. There is a separate analysis with only vegans. This info isn't in the main article, but this info can be found in an additional PDF, in "supplementary table 3" (see below). Vegans consumed more soya foods, fruits, vegetables and nuts than non-vegan vegetarians (Tong et al. 2019 supplementary PDF). --- This is good, and should also be good regarding stroke prevention.

You can ignore this ...
"When we assessed vegetarians and vegans separately, the point estimates for vegans were lower for ischaemic heart disease (0.82, 0.64 to 1.05) and higher for total stroke (1.35, 0.95 to 1.92) than meat eaters, but neither estimate was statistically significant, possibly because of the small number of cases in vegans [low total number of vegans with stroke], as indicated by the wide confidence intervals (supplementary table 3)." (Tong et al. 2019)
"The associations were attenuated after adjustment for self reported high blood pressure, high blood cholesterol, diabetes, and body mass index, which suggests that part of the associations might be attributed to these factors. However, the lower risk in vegetarians and vegans remained marginally significant after adjustment for all of these factors. The reason for such differences is not certain, but could be partly attributed to lower concentrations of low density lipoprotein cholesterol (LDL-C; or non-HDL-C concentrations as a surrogate) associated with meatfree diets,29-31 differences that were not fully accounted for by adjusting for self reported high blood cholesterol." (Tong et al. 2019)

... and skip right to Supplementary table 3: ("Translation" below)
In the left column there are five sections, in plain English: (1) heart attack, (2) heart disease caused by a blocked artery, (3) stroked caused by a blocked artery, (4) stroke caused by bleeding (haemorrhage), (5) all types of stroke (including the previous two types).

Compared to the meat eater group ... 
(1) Vegans had a 23% reduced risk of heart attack - better than all other groups - BUT THIS WAS NOT statistically significant. The confidence interval for vegans (these numbers "(0.46 to 1.27)") was much wider than for the other groups. This could mean that there were some vegans with a much lower risk (possibly those who got enough vitamin B12), and some with a much higher risk (possibly those who were quite B12-deficient). Interestingly, the fish eaters (pescatarians) had a 0% reduced risk of heart attack. For non-vegan vegetarians the results (9% lower risk) was not significant either.
(2) Vegans had a 18% reduced risk of heart problems (caused by artery blockage, including heart attack) - BUT THIS WAS NOT statistically significant. On the other had fish eaters had a 13% and non-vegan vegetarians had a 23% lower risk which WAS significant. Again, the reason why vegans did not have a clearly lower risk might be that some or many of these vegans were B12-deficient.
(3) THIS I THINK IS THE MOST IMPORTANT NUMBER: Vegans had a 54% increased risk of stroked caused by artery blockage - but again THIS WAS NOT significant. Again, it might be that B12-deficiency in many vegans caused this result - it wasn't significant, but still the numbers don't look the way we would like them to look. Results for fish eaters were a 5% higher risk, and for non-vegan vegetarians a 7% higher risk, BUT THESE RESULTS WERE NOT significant. 
(4) Vegans had a 9% higher risk for stroke caused by bleeding, BUT it was NOT significant. Again, the confidence interval for vegans (these numbers: "(0.53 to 2.26)") was very wide, indicating that vegans in general COULD even have a lower risk (or a much higher risk). In any case, vegans should make sure to get enough vitamin B12 and enough omega-3 fatty acids, while it's probably also a good idea to avoid a great excess of omega-3s, as this can cause bleeding (small bruises) at least in some people. For fish eaters their 12% higher risk was NOT significant either. But non-vegan vegetarians had a 48% higher risk which WAS statistically significant! THIS is the weird news that the media are talking about. Now, the only "bad thing" that these vegetarians consumed more than the meat eaters was cheese: on average 30 grams per day (vs. meat eaters who ate 21 g/d) - but this quite surely isn't the reason. The reason could be that many of these vegetarians were also B12-deficient (wild speculation, see Vanoli et al. 2018 and Masoodi et al. 2011). Probably, some vegetarians ate lots of dairy, while others were near-vegan (Tong et al. 2019 supplementary PDF) while at the same time not consuming enough B12. 
(5) THIS NUMBER AGAIN IS INTERESTING, because it includes all types of stroke, including haemorrhagic (bleeding) and ischaemic (artery blockage) stroke, but also any other type. Vegans had a 35% higher risk, BUT THIS WAS NOT significant. Fish eaters had a 14% higher risk and non-vegan vegetarians had a 17% higher risk - both NON-significant. Again, there is nothing else that seems as obvious and that one could speculate about: The reason for this higher tendency for stroke (yes, NOT significant) could be B12 deficiency in many vegans.
--- So if there is anything to these numbers, unlike with non-vegan vegetarians, for vegans it seems that ischaemic stroke (artery blockage) is more of an issue that haemorrhagic stroke (bleeding) - but the reason for increased "artery blockage" in the brain region (causing stroke) and possibly even for haemorrhagic stroke (bleeding) might be B12 deficiency.   

The hypothesis that B12 deficiency could be the culprit for these non-significant results of the vegans in this study having a higher risk of stroke than meat eaters - and why the heart attack risk in these vegans wasn't as amazing as we'd like it to be ... is backed up by two findings:
- Of the 1832 vegans included in this study only 51% (my calculation) or 52% (their calculation) took any kind of supplement (such as "any vitamins, minerals, [..] fibre or other food supplements during the last 12 months"). So, unless the other half of the vegans was consuming enough B12 from fortified foods they could be expected to be B12-deficient.
- This is exactly what a smaller and older analysis from the same study (EPIC-Oxford) showed: Vegans on average had less B12 in their blood, i.e. many were B12-deficient
"Vegetarians and vegans in the EPIC-Oxford cohort have lower circulating levels of several nutrients (eg, vitamin B12, vitamin D, essential amino acids, and long chain n-3 polyunsaturated fatty acids), and differences in some of these nutritional factors could contribute to the observed associations." (Tong et al. 2019) [For nutrient recommendations for vegans, see here]

More numbers:
Total number of study participants: 48188
Total number of vegan study participants: 1832 vegans (1169 vegan women (64%), and 663 vegan men (36%))
Average age of the vegans: 38.6 years (standard deviation: 13.2)
Number of vegans who took a dietary supplement: 935 (52.1%; I would say 935/1832 =~51.0%)

This finding about higher stroke risk in vegans in the EPIC-Oxford study IS NOT THAT NEW. An analysis from EPIC-Oxford from 2016 which looked at "death from stroke" (i.e. cases of fatal strokes, not just cases of all strokes) found that vegans had a ~50% higher risk, but this was not statistically significant (see here, under EPIC-Oxford (2016)). HOWEVER, a later analysis that included not just EPIC-Oxford (from the UK) but also a bigger study with more vegans (AHS2; from the USA) found that vegans had a 7% lower risk of death from stroke compared to meat eaters which again was NOT statistically significant (see here, under Meta-analysis by Dinu et al. (2017)).

An editorial in the same journal as the above mentioned article by Tong et al. 2019 also discussed the possible reasons for these results:
"[RE: Tong et al. (2019):] Conversely, the study also showed that vegetarians had a 20% higher risk of total stroke than meat eaters (hazard ratio 1.20 (95% confidence interval 1.02 to 1.40)), mostly due to a higher rate of haemorrhagic stroke. Fish eaters had a non-significant 14% higher rate of total stroke than meat eaters (1.14 (0.94 to 1.38)).
The higher relative risk of stroke among vegetarians is a new contribution to the body of evidence on the health effects of a vegetarian diet. Tong and colleagues’ study has many strengths that diminish the likelihood that this association is an artefact.
Vegetarians and others should keep the reported stroke risk in perspective, however. It is based on results from just one study, and the increase is modest relative to meat eaters: “equivalent to three more cases of total stroke (95% confidence interval 0.8 to 5.4 more) per 1000 population over 10 years.” Relevance to vegetarians worldwide must also be considered. Participants were all from the United Kingdom where dietary patterns and other lifestyle behaviours are likely to differ from those prevalent in low and middle income countries, where most of the world’s vegetarians live.
Vitamin B12 is considered a nutrient at risk in some vegetarian diets, unless fortified foods and supplements are used. The role of suboptimal intake of B12 in stroke risk is unclear, and further exploration should include re-evaluation of existing vitamin B trials and mechanistic studies to support observational evidence.
[...] populations in some low and middle income countries who consume very low amounts of animal source foods may benefit from being able to eat a little more of these foods [or consume supplements or fortified foods] to gain additional nutrients [especially vitamin B12, but also iodine] necessary to help combat all forms of malnutrition." (Lawrence and McNaughton 2019)

Even though, this editorial states that the causal relation between B12 deficiency and higher stroke risk is unclear - everything will always be somewhat unclear - a 2019 review about diet and stroke in the journal Nutrients by stroke expert J. David Spence states:

"Nutrition is far more important in stroke risk than most physcians [physicians] suppose. Healthy lifestyle choices reduce the risk of stroke by ~80%, and of the factors that increase the risk of stroke, the worst is diet: only ~0.1% of Americans consume a healthy diet, and only 8.3% consume a somewhat healthy diet. The situation is probably not much better in most other countries. A [traditional] Cretan Mediterranean diet [which is plant-based], high in olive oil, whole grains, fruits, vegetables and legumes, and low in cholesterol and saturated fat, can reduce stroke by 40% or more in high-risk patients [THIS TYPE OF DIETARY PATTERN CAN BE VEGAN, OF COURSE]. [...] Metabolic B12 deficiency is common [in older people] and usually missed. It has serious neurological consequences, including an increase in the risk of stroke. It now clear that B vitamins to lower homocysteine reduce the risk of stroke, but we should probably be using methylcobalamin instead of cyanocobalamin [in people with kidney disease] [...].
Patients at risk of stroke should keep their salt intake to 2–3 grams per day, learn to make a Mediterranean diet tasty and enjoyable, should limit their intake of meat (particularly red meat) and should avoid egg yolk. Such changes represent a challenge for many patients, so physicians should educate patients on the importance of these measures, and provide them with assistance. They should also have metabolic B12 deficiency and hyperhomocysteinemia detected and treated, probably with methylcobalamin [or any type of B12 if you have healthy kidneys]. Such measures have the potential to greatly reduce the risk of stroke, so deserve greater attention from the public, and particularly from physicians." (Spence 2019)

Several news reports have documented a statement by the here discussed study's first author Dr Tammy Tong "Recent evidence suggests that very low cholesterol levels might be linked to a higher risk of hemorrhagic stroke, Tong noted. Vegetarians and vegans might also have low levels of some nutrients, such as vitamin B12 [...]". (US NEWS 5 September 2019)
As mentioned above, the haemorrhagic stroke risk in this study - for unknown reasons - was significantly higher in non-vegan vegetarians, but not in vegans. So, if we only look at vegans, I would tend to look more at the 54% increased ischaemic (artery blockage) stroke risk.
But, looking at haemorrhagic stroke for a moment: It seems bizarre but true (there is evidence) that lower cholesterol levels - which are considered desirable and which are recommended for cardiovascular disease prevention worldwide - seem to somewhat increase the risk of brain bleeding (Xie et al. 2017, Chen et al. 2017, Teoh et al. 2019).

If you've had a cholesterol test (or two) done and your cholesterol is extremely low, and you also don't feel well, try eating some coconut oil, and a lot more fat in general (healthy fats such as nuts, seeds, nut butters, avocadoes, olives, olive oil, etc.). This is uncommon in vegans.
However, if you generally feel well and your cholesterol is in the officially recommended area, it cannot really be recommended that you should try to raise your cholesterol levels (and therefore your risk of heart attack and stroke), just to potentially lower your relatively small risk of one subtype of stroke (Sun et al. 2019).

"Funnily" very low cholesterol could also make some susceptible individuals more aggressive, especially men with "mental health issues": "Men seem to be more sensitive to low cholesterol levels as the association between low cholesterol levels and aggression is found mostly in men. [...] Lowering cholesterol levels with statins brings about several changes in the serotonergic system, nerve cell membrane microviscosity and behaviour, and needs to be done with precaution in susceptible individuals. Cholesterol levels could serve as a biological risk marker for violence and suicidal tendencies in psychiatric patients with depression and schizophrenia." (Tomson-Johanson and Harro et al 2018)

But ... isn't this just another study sponsored by the meat industry?
No, it's not. As described above, EPIC-Oxford is one of the two studies in the world which includes a large number of vegans. In fact, two of the authors of the study are vegans, and they are two of the most experienced and widely respected nutrition scientists in the word (Tong et al. 2019): Tim Key (the last author - "last author" means "big boss" of the study), and Paul Appleby. They are both - I assume - members of the Vegan Society. And this is not a secret - and as you can expect they are regularly attacked by the "emotional wannabe-science internet people" as being biased because they are vegan. As if being a vegan automatically made you biased, and as if eating animals automatically made you oh so neutral when it comes to the health effects of eating animals.

As vegans we are a small minority and we are constantly attacked the "the whole world". Let's not attack people who have dedicated their lives to finding out what benefits vegan diets can have for people's health and the environment, and also what risks there could be, and how we can avoid those risks.

I understand that as a vegan - and the average vegan is not a nutrition scientist and doesn't have to be (actually it might be quite unhealthy to be a researcher, lots of sitting and researching, and stress) - you might feel a bit queasy when you see such news, and the barrage of mainstream carnivore media articles and comments seemingly saying "haha veganism is unhealthy, we always knew it".

You might feel like questioning your whole existence (exaggeration) ...

"Can it be that there's some sort of error
Hard to stop the surmounting terror
Is it really the end, not some crazy dream?"

Rest assured that a vegan diet is safe - but you really should take a B12 supplement (or eat plenty of fortified foods or use a B12-fortified tooth paste). 
Like Propagandhi used to say (they're probably still saying it): Knowledge is a weapon, arm yourself. So, arm yourself with some knowledge: We do not need meat, and we do not need eggs or dairy. We know this, and this is not really controversial (see here).   
A healthy vegan diets contains plenty of fruit and vegetables, and nuts, whole grains and legumes, and you should consume enough B12 and ideally know a little bit (not much) about a few more nutrients and how you can consume them in optimal amounts: Nutrient recommendations for vegans     

If you feel secure in your knowledge you can sit back (and jump up) and relax ... and eat some falafel maybe.

And relax, because stress can kill you.

Update 15 October 2019:
Oral contraceptive use might increase the risk of stroke in women - for ischaemic and haemorrhagic stroke and all types of stroke combined (Li et al. 2019).

The use of antidepressant medication might also increase the risk of stroke (Trajkova et al. 2019).

Higher consumption of meat, salty snacks and fried foods, and lower consumption of fruit and green leafy, cooked or raw vegetables might be associated with a higher risk of both ischaemic and haemorrhagic stroke (Chen et al. 2019).

Alcohol consumption increases the risk of haemorrhagic stroke. A recent analysis found that in current drinkers there is a 17% increased risk of haemorrhagic stroke per 100 g
Hazard ratios for subtypes of cardiovascular outcomes in current drinkers, per 100 g per week "higher usual alcohol consumption" (Wood et al. 2018). 100 g are about 1.25 litres of wine or 3 litres of beer.

Tobacco smoking and passive smoking increase the risk of stroke (Pan et al. 2019).

Update 5 March 2020:
Taking vitamin E supplements or vitamin E-containing might increase the risk of haemorrhagic stroke (Cortés-Jofré et al. 2020).

Update 4 November 2020:
"Our results showed that Hcy [homocysteine, which increases with B12 deficiency] levels in ICH [intracerebral hemorrhage] patients were significantly higher than those in healthy controls (SMD = 0.59, 95% CI = 0.51-0.68, P < 0.001) [this means 59% higher risk] [...] In conclusion, Hcy level may be an aggravating factor in atherosclerosis, which is positively associated with high risk of ICH." (Zhou et al. 2018)

Physical activity/exercise is also good for stroke prevention (Prior and Suskin 2018).

Update 1 August 2023 regarding EPA/DHA:
"Of note, neither VITAL nor the majority of the first- or second-wave secondary prevention trials [the largest randomized controlled trials on this topic] found a benefit [of EPA/DHA supplementation] for stroke (the strong reduction in REDUCE-IT is one exception) […]. Interestingly, observational studies also show more consistent effects of n−3 Fas [EPA/DHA] on coronary outcomes than on stroke. As noted earlier, high dietary intakes and high blood levels of n−3 FAs [EPA/DHA] both are associated with reduced coronary risk in observational studies. In contrast, high fish intakes [fish can be good sources of vitamin B12, vitamin D, iodine, selenium, iron, zinc, protein – plus fish is often eaten instead of meat] are more consistently inversely associated with ischaemic stroke risk than are high levels of n−3 FA biomarkers." (Bassuck et al. 2023)

Update 25 February 2024:
Drinking green tea regularly also appears to reduce strike risk (Wang et al. 2023, Deng et al. 2024).